A new scheme for energy-efficient estimation in a sensor network
Xun Chen, Rick S. Blum, Brian M. Sadler · 2009
In this paper, energy efficient estimation of an unknown parameter in Gaussian noise is studied in a sensor networking context. A new approach is suggested to obtain a good approximation to the traditional maximum likelihood (ML) estimate, which can save energy by reducing the number of sensor transmissions. Specifically, we describe a new and simple transmission scheme in which the sensor transmissions are ordered according to the magnitude of their measurements, and the sensors with small magnitude measurements, smaller than a threshold, do not transmit. A bound on the error of approximation is derived, which can be utilized to dynamically determine the threshold such that a trade-off between the accuracy of the approximation and the energy savings can be maintained. Through the numerical results, we show that our approach can be very energy efficient with only a negligible estimation error introduced.